Background: In Spring 2014 the VA launched a national Opioid Overdose Education and Naloxone Distribution (OEND) program with the support of the VA Under Secretary for Health1. In launching VA's OEND initiative, the Under Secretary for Health acknowledged that most of the evidence to date surrounds OEND targeting community-dwelling persons with documented opioid use disorders or opioid misuse rather than all patients receiving opioid analgesics in a health care system. The VA OEND program was the first to target two distinct patient populations: (1) patients with opioid use disorders and (2) patients prescribed opioid analgesics. As of May 22, 2016 VA facilities have dispensed just over 27,000 naloxone kits to Veterans with every VA facility participating in naloxone distribution (VA Naloxone Kit Distribution Report). As the first large health care system in the United States to implement OEND, there is a great deal that can be learned from VA in terms of whether current recommendations for OEND implementation are sound and/or whether they should be tailored to specific subsets of patients. To address concerns about the effectiveness of OEND in health care settings, an evaluation of OEND is necessary to examine whether it reduces opioid overdose as intended for at-risk patients prescribed opioid analgesics and patients with opioid use disorders. Objectives: The proposed study will pursue the following objectives: Objective 1: Characterize naloxone distribution within VA and patient-, prescriber-, and setting-related factors associated with distribution. Objective 2: Assess whether naloxone distribution to at-risk Veterans compared to similar at-risk Veterans who did not receive naloxone is associated with reduced fatal and non-fatal opioid overdose. Theoretical Framework: The Reach, Effectiveness, Adoption, Implementation, and Maintenance?RE-AIM? framework will guide our study. This framework considers both individual- (e.g., patient) and institutional-level (e.g., setting) factors when evaluating the impact of interventions. Methods: We will identify a cohort of patients from Fiscal Year 2014 Quarter 3 (FY2014 Q3; the start of the national OEND program) through FY2017 Q1 (end of calendar year 2016) comprising at-risk patient populations: (1) patients with opioid use disorders and (2) at-risk patients prescribed opioid analgesics. Objective 1: For the identified cohort, we will obtain data from the VA Corporate Data Warehouse (CDW) and provide descriptive statistics on the characteristics of patients who receive naloxone, prescribers who prescribe naloxone, and settings in which naloxone is prescribed. We will also use mixed-effects logistic regression models to identify patient-, prescriber-, and setting-level differences between patients who do and do not receive naloxone. Objective 2: Our primary analysis will use marginal structural models (MSMs) to examine the effect of naloxone distribution for two main sets of outcomes: (1) Fatal opioid overdose?unintentional, intentional, and combined and (2) Non-fatal opioid overdose?unintentional, intentional, and combined. MSMs are a method for controlling for selection bias by using inverse probability of treatment weights which follow many of the same principles as propensity score analysis. Uncommon observations (persons who get the unexpected treatment) get greater weight, and common observations get less weight. All at-risk patients will be included in our analyses and we will obtain fatal overdose outcomes from the VA/Department of Defense (DoD) National Death Index and non-fatal overdose outcomes from CDW. Summary: The VA is leading the nation with regards to health care system-based implementation of naloxone distribution. The proposed study will characterize implementation of this innovative program and assess whether naloxone distribution is meeting its intended goal of reducing opioid overdose among at-risk Veterans.